This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
A 3-month videoconference interaction program with family members has been shown to decrease depression and loneliness in nursing home residents. However, little is known about the long-term effects on residents’ depressive symptoms, social support, and loneliness.
The purpose of this longitudinal quasi-experimental study was to evaluate the long-term effectiveness of a videoconference intervention in improving nursing home residents’ social support, loneliness, and depressive status over 1 year.
We purposively sampled 16 nursing homes in various areas of Taiwan. Elderly residents (N = 90) of these nursing homes meeting our inclusion criteria were divided into an experimental (n = 40) and a comparison (n = 50) group. The experimental group received at least 5 minutes/week for 3 months of videoconference interaction with their family members in addition to usual family visits, and the comparison group received regular family visits only. Data were collected in face-to face interviews on social support, loneliness, and depressive status using the Social Support Behaviors Scale, University of California Los Angeles Loneliness Scale, and Geriatric Depression Scale, respectively, at four times (baseline, 3 months, 6 months, and 12 months after baseline). Data were analyzed using the generalized estimating equation approach.
After the videoconferencing program, participants in the experimental group had significantly lower mean change in instrumental social support scores at 6 months (–0.42,
Our videoconference program had a long-term effect in alleviating depressive symptoms and loneliness for elderly residents in nursing homes. This intervention also improved long-term emotional social support and short-term appraisal support, and decreased residents’ instrumental social support. However, this intervention had no effect on informational social support.
Similar to other countries, Taiwan has more and more older people living in nursing homes. The number of nursing homes in Taiwan grew from 10 in 1995 to 372 in 2010 [
Among these interventions, social support is one of top importance because the social support systems older people use are closely related, both in quality and in quantity, to their health and quality of life [
Although most nursing home residents have become functionally dependent due to poor physical health, their psychosocial needs do not decrease [
The benefits of videoconferencing in medicine have been recognized as a feasible way of delivering care to frail older people living with chronic diseases [
To date, no empirical studies have used a longitudinal design to examine the effectiveness of videoconferencing for nursing home residents in Taiwan. Understanding the effectiveness of videoconferencing in Taiwanese nursing homes would fill the knowledge gap on this topic. Thus, the purpose of this study was to evaluate the long-term follow-up effectiveness of a videoconference intervention program on nursing home residents’ social support, loneliness, and depressive status after a minimum 3-month videoconference program.
A quasi-experimental longitudinal design was used. Nursing homes from northern and central Taiwan were purposively selected based on three criteria: size (>65 beds), with Internet access, and accessible to the researchers. Among 23 medium-large nursing homes that met the recruitment criteria, 7 declined to participate in our study. The remaining 16 nursing homes (total beds =2190) were therefore randomly assigned to the comparison or experimental group.
The sample size was estimated based on the rule that 15 participants were needed for each variable [
The videoconference program used laptops and Internet communication programs. Nursing home residents were asked to use the Internet at least once a week, with help from a trained research assistant, who spent at least 5 minutes per week with each resident for the first 3 months during their scheduled videoconference visit. This weekly frequency was chosen to reflect the frequency of in-person visits to a nursing home resident for the majority of families [
The study was approved by the Institutional Review Board of the authors’ institution. After permission was granted from the nursing homes’ administrators, details of our research procedure were posted at the entrance of each nursing home. This announcement indicated that residents or family members interested in participating in the study could directly contact the study personnel or nursing home staff. We also asked nursing home staff to talk with residents who met our study criteria and their family members about their willingness to participate in this study. Those who were interested in participation were contacted by the research assistant, who explained their rights, benefits, confidentiality, and responsibilities when participating in the study. After signing informed consent, residents and family members made appointments for videoconferences. Family members were phoned or emailed before the scheduled time to remind them of the appointment. The family of residents in the experimental group could continue their in-person or telephone visits as usual. Laptops were left in the nursing homes for 1 year. For the first 3 months, residents were helped by a trained research assistant to use the videoconference technology in a private room; for the next 9 months, residents were helped by trained nursing home staff. All residents in both the experimental and comparison groups completed questionnaires for demographic information (baseline only), depressive symptoms, loneliness, and social support at baseline and at 3, 6, 9, and 12 months.
Demographic indicators included participants’ age, gender, marital status, educational background, duration of residency in the nursing home, and frequency of family visits. Residents’ physical status and cognitive status were measured at baseline using the Barthel Index [
Depressive status was measured by the Geriatric Depression Scale (GDS) [
Loneliness was measured by the revised University of California Los Angeles (UCLA) Loneliness Scale [
Social support was measured by the Social Support Behaviors Scale with three subscales: social support network, quantity of social support, and satisfaction with social support [
Family involvement with residents was confirmed by asking nursing home staff to record the number of visits and phone calls made to the residents. The duration of each videoconference interaction during the year was recorded by either the research assistant or nursing staff. Videoconference use was coded as the frequency of all videoconference interactions per month.
All data were coded before being entered into a computer. Statistical analyses were performed using SPSS for Windows version 15.0 (IBM Corporation, Somers, NY, USA). Participants’ demographic data were analyzed by descriptive statistics. Differences between groups were compared at four points (baseline, 3 months, 6 months, and 12 months) using multiple linear regression with the generalized estimating equation (GEE) method [
The 40 participants in the experimental group were on average 73.82 (SD 11.19) years old at baseline. The experimental group’s use of videoconferencing decreased over time at 3, 6, and 12 months: mean (SD) 2.09 (1.46), 1.69 (1.37), and 1.14 (1.22), respectively. However, this decrease was not statistically significant. The majority of participants were female (22/40, 55%) and widowed (29/40, 73%), and 35% (14/40) had graduated from primary school. Their average MMSE and Barthel Index scores at baseline were 23.51 (SD 3.93) and 65.68 (SD 22.55), respectively, indicating good cognitive status and above average performance of ADLs. They had an average of 3.69 (SD 2.09) children. About half of these participants (22/40, 55%) were visited by a family member at least once a week, and only 18% (7/40) seldom (less than once a month) had a family member visit them. Their average length of residency was 28.38 (SD 30.79) months (
Demographic characteristics of experimental and comparison groups
Variable | Comparison group (n = 50) | Experimental group (n = 40) | χ2(dfa; |
|
|||
n (%) | Mean (SD) | n (%) | Mean (SD) | ||||
Age (years) | 79.26 (7.07) | 73.82 (11.19) | 10.78 (.01) | ||||
|
0.2 (1; .63) | ||||||
Male | 20 (40) | 18 (45) | |||||
Female | 30 (60) | 22 (55) | |||||
|
6.6 (3; .16) | ||||||
Single | 1 (2) | 2 (5) | |||||
Married | 20 (40) | 8 (20) | |||||
Divorced | 3 (6) | 1 (2) | |||||
Widowed | 26 (52) | 29 (73) | |||||
|
12.7 (4; .06) | ||||||
None/illiterate | 29 (58) | 9 (23) | |||||
Primary | 8 (16) | 14 (35) | |||||
Junior high school | 2 (4) | 2 (5) | |||||
Senior high school | 8 (16) | 11 (28) | |||||
≥ College | 3 (6) | 4 (10) | |||||
Number of children | 3.90 (2.05) | 3.69 (2.09) | 0.10 (.64) | ||||
Residency (months) | 29.32 (28.58) | 28.38 (30.79) | 0.04 (.87) | ||||
Barthel Index | 63.10 (23.62) | 65.68 (22.55) | 0.16 (.61) | ||||
MMSE | 22.22 (3.93) | 23.51 (3.93) | 0.05 (.13) | ||||
|
|||||||
None/seldom | 7 (14) | 7 (18) | |||||
Monthly | 5 (10) | 11 (28) | |||||
Weekly (>2/month) | 34 (68) | 18 (45) | |||||
Daily (>5/week) | 4 (8) | 4 (10) | |||||
|
|||||||
0 | 36 (72) | 24 (60) | |||||
1 | 10 (20) | 8 (20) | |||||
2–6 | 3 (6) | 6 (15) | |||||
≥7 | 1 (2) | 2 (5) |
a Degrees of freedom.
The 50 participants in the comparison group were on average 79.26 (SD 7.07) years old at baseline. The majority were female (30/50, 60%), had no formal education (29/50, 58%), and were widowed (26/50, 52%). Their average MMSE and Barthel Index scores were 22.22 (SD 3.93) and 63.10 (SD 23.62), respectively, indicating good cognitive status and above average performance of ADLs. Their average number of children was 3.90 (SD 2.05). Most participants (38/50, 76%) were visited by a family member at least once a week and 14% (7/50) seldom had a family member visit them. Their average length of residency was 29.32 (SD 28.58) months (
During the 12-month study, 13 participants in the experimental group withdrew from the study (including 5 who declined to continue participating, 2 who relocated back home, and 3 who died), with an attrition rate of 33%. In the comparison group, 22 participants dropped out (including 5 who died, 12 who relocated back home, and 4 who developed cognitive deficits), with an attrition rate of 44% (
Attrition of participants in the two groups over the 1-year study period.
Descriptive analysis of outcomes shows that both groups had the highest social support scores for the informational and instrumental social support subscales. The mean GDS scores for depressive status at baseline, and 3, 6, and 12 months were 12.75, 11.57, 12.85, and 13.00, respectively, for the experimental group, and 10.52, 10.56, 14.41, and 15.04, respectively, for the comparison group. These GDS scores did not differ significantly by independent
Social support (Social Support Behaviors Scale), depressive status (Geriatric Depression Scale), and loneliness (University of California Los Angeles Loneliness Scale) scores by group at baseline, and 3, 6, and 12 months
Variable | Comparison group (n = 50) | Experimental group (n = 40) |
|
|||
Mean | SD | Mean | SD | |||
|
||||||
Baseline | 9.76 | 1.69 | 9.49 | 1.57 | 0.78 (.59) | |
3 months | 9.40 | 1.55 | 9.71 | 1.72 | –0.80 (.66) | |
6 months | 9.24 | 1.28 | 9.30 | 1.67 | –0.14 (.24) | |
12 months | 8.96 | 0.98 | 9.31 | 1.80 | –2.30 (.04) | |
|
||||||
Baseline | 10.66 | 1.82 | 11.13 | 1.48 | –1.31 (.21) | |
3 months | 10.62 | 1.20 | 11.31 | 1.41 | –2.33 (.02) | |
6 months | 10.60 | 1.55 | 11.02 | 1.43 | –1.08 (.61) | |
12 months | 10.28 | 1.32 | 10.76 | 1.63 | –1.16 (.25) | |
|
||||||
Baseline | 10.40 | 1.05 | 10.50 | 1.22 | –0.38 (.33) | |
3 months | 10.36 | 0.84 | 10.35 | 1.26 | 0.05 (.95) | |
6 months | 10.53 | 1.05 | 10.31 | 1.14 | 0.77 (.51) | |
12 months | 10.13 | 0.83 | 10.01 | 1.17 | 0.45 (.12) | |
|
||||||
Baseline | 9.25 | 1.60 | 9.03 | 1.25 | 0.73 (.14) | |
3 months | 8.83 | 1.26 | 9.30 | 1.32 | –1.59 (.12) | |
6 months | 8.98 | 1.28 | 9.30 | 1.58 | –0.87 (.39) | |
12 months | 8.70 | 1.22 | 9.26 | 1.67 | –1.42 (.16) | |
|
||||||
Baseline | 141.54 | 18.64 | 141.85 | 17.26 | –0.08 (.93) | |
3 months | 139.00 | 14.95 | 143.11 | 18.36 | –1.07 (.28) | |
6 months | 139.55 | 16.08 | 140.43 | 18.02 | –0.20 (.63) | |
12 months | 134.08 | 12.74 | 138.08 | 19.36 | –0.85 (.40) | |
|
||||||
Baseline | 10.52 | 4.06 | 12.75 | 5.50 | 1.94 (.06) | |
3 months | 10.56 | 3.89 | 11.57 | 5.27 | –0.96 (.34) | |
6 months | 14.41 | 4.93 | 12.85 | 5.35 | 1.16 (.25) | |
12 months | 15.04 | 4.61 | 13.00 | 4.50 | 1.60 (.11) | |
|
||||||
Baseline | 45.76 | 9.73 | 49.70 | 10.25 | –1.85 (.09) | |
3 months | 45.59 | 9.40 | 44.54 | 12.68 | 0.41 (.68) | |
6 months | 47.81 | 9.97 | 46.21 | 11.87 | 0.57 (.56) | |
12 months | 48.32 | 10.17 | 45.92 | 12.14 | 0.78 (.47) |
a Degrees of freedom.
Time effects between the two groups were compared using the GEE method’s multiple linear regression. As shown in
Effects of videoconference intervention on participants’ depressive status and loneliness at 3, 6, and 12 months in consideration of group, time, and time × group effects
Variable | Unadjusted | Adjusteda | |||||||||
beta | SE | χ2 1 |
|
beta | SE | χ2 1 |
|
||||
|
|||||||||||
|
|||||||||||
E vs Cb | 2.22 | 1.03 | 5.0 | .06 | 2.26 | 1.00 | 5.1 | .05 | |||
|
|||||||||||
3 months | 0.97 | 0.33 | 8.9 | <.001 | 0.99 | 0.32 | 9.3 | <.001 | |||
6 months | 6.95 | 0.63 | 123.4 | <.001 | 6.99 | 0.63 | 142.7 | <.001 | |||
12 months | 7.64 | 0.70 | 118.9 | <.001 | 7.71 | 0.70 | 119.6 | <.001 | |||
|
|||||||||||
3 months | –1.36 | 0.56 | 6.0 | .02 | –2.64 | 0.57 | 21.3 | <.001 | |||
6 months | –4.50 | 0.97 | 21.6 | <.001 | –4.33 | 1.03 | 17.6 | <.001 | |||
12 months | –4.45 | 0.89 | 24.9 | <.001 | –4.40 | 0.92 | 23.1 | <.001 | |||
|
|||||||||||
|
|||||||||||
E vs Cb | 3.94 | 2.10 | 3.5 | .09 | 3.27 | 2.23 | 2.2 | .14 | |||
|
|||||||||||
3 months | –0.31 | 0.53 | 0.4 | .55 | –0.31 | 0.53 | 0.3 | .56 | |||
6 months | 2.81 | 1.23 | 5.2 | .02 | 2.81 | 1.23 | 5.2 | .02 | |||
12 months | 2.77 | 1.22 | 5.2 | .02 | 2.78 | 1.23 | 5.1 | .02 | |||
|
|||||||||||
3 months | –4.84 | 1.14 | 18.0 | <.001 | –5.40 | 1.22 | 19.6 | <.001 | |||
6 months | –6.46 | 1.64 | 15.4 | <.001 | –6.47 | 1.70 | 14.5 | <.001 | |||
12 months | –6.42 | 1.64 | 15.3 | <.001 | –6.27 | 1.94 | 10.5 | .001 |
a Adjusted for residents’ age and length of residency.
b C: comparison group, E: experimental group.
c B: baseline measurement.
d Group 0: comparison group (reference group), group 1: experimental group.
This study had nine outcome variables of interest (depressive status, loneliness, total social support, emotional support, informational support, instrumental support, appraisal support, number of phone calls, and number of visits). Each outcome variable was analyzed exactly as in
Effects of videoconference intervention on participants’ social support, depressive status, and loneliness at 3, 6, and 12 months in consideration of time × group effects
Variable | Unadjusted | Adjusteda | ||||||||
beta | SE | χ2 1 |
|
beta | SE | χ2 1 |
|
|||
|
||||||||||
3 months | 3.71 | 2.10 | 3.1 | .08 | 3.63 | 2.10 | 3.0 | .09 | ||
6 months | –0.58 | 2.50 | 0.1 | .82 | –0.84 | 2.52 | 0.1 | .74 | ||
12 months | –0.05 | 2.73 | 0.0 | .99 | –0.48 | 2.74 | 0.0 | .86 | ||
|
||||||||||
3 months | 0.60 | 0.19 | 10.1 | .001 | 0.74 | 0.19 | 15.3 | <.001 | ||
6 months | 0.32 | 0.26 | 1.5 | .23 | 0.40 | 0.28 | 2.1 | .15 | ||
12 months | 0.47 | 0.25 | 3.5 | .06 | 0.61 | 0.26 | 5.3 | .02 | ||
|
||||||||||
3 months | –0.00 | 0.23 | 0.0 | >.99 | 0.15 | 0.25 | 0.4 | .53 | ||
6 months | –0.29 | 0.27 | 1.2 | .28 | –0.24 | 0.31 | 0.6 | .44 | ||
12 months | –0.34 | 0.31 | 1.2 | .28 | –0.18 | 0.36 | 0.3 | .62 | ||
|
||||||||||
3 months | –0.20 | 0.14 | 2.0 | .16 | –0.14 | 0.15 | 0.9 | .34 | ||
6 months | –0.47 | 0.18 | 6.7 | .01 | –0.42 | 0.19 | 4.7 | .03 | ||
12 months | –0.41 | 0.19 | 4.9 | .03 | –0.41 | 0.19 | 4.6 | .03 | ||
|
||||||||||
3 months | 0.66 | 0.22 | 9.3 | .002 | 0.74 | 0.22 | 10.8 | .001 | ||
6 months | 0.37 | 0.24 | 2.3 | .13 | 0.43 | 0.26 | 2.8 | .10 | ||
12 months | 0.57 | 0.31 | 3.5 | .06 | 0.58 | 0.32 | 3.3 | .07 | ||
|
||||||||||
3 months | –1.36 | 0.56 | 6.0 | .02 | –2.64 | 0.57 | 21.3 | <.001 | ||
6 months | –4.50 | 0.97 | 21.6 | <.001 | –4.33 | 1.03 | 17.6 | <.001 | ||
12 months | –4.45 | 0.89 | 24.9 | <.001 | –4.40 | 0.92 | 23.1 | <.001 | ||
|
||||||||||
3 months | –4.84 | 1.14 | 18.0 | <.001 | –5.40 | 1.22 | 19.6 | <.001 | ||
6 months | –6.46 | 1.64 | 15.4 | <.001 | –6.47 | 1.70 | 14.5 | <.001 | ||
12 months | –6.42 | 1.64 | 15.3 | <.001 | –6.27 | 1.94 | 10.5 | .001 | ||
|
||||||||||
3 months | 0.28 | 0.11 | 6.3 | .01 | 0.28 | 0.12 | 5.8 | .02 | ||
6 months | 0.22 | 0.14 | 2.5 | .11 | 0.20 | 0.15 | 1.7 | .19 | ||
12 months | –0.01 | 0.15 | 0.0 | .97 | 0.01 | 0.16 | 0.0 | .95 | ||
|
||||||||||
3 months | –0.08 | 0.07 | 1.6 | .21 | –0.08 | 0.07 | 1.2 | .27 | ||
6 months | 0.05 | 0.08 | 0.4 | .53 | 0.03 | 0.08 | 0.2 | .70 | ||
12 months | 0.05 | 0.08 | 0.4 | .55 | 0.04 | 0.08 | 0.2 | .67 |
a Adjusted for residents’ age and length of residency.
b Group 0: comparison group (reference group), group 1: experimental group.
This study demonstrated that our videoconference intervention alleviated elderly nursing home residents’ perceived loneliness and improved their depressive status at 3, 6, and 12 months after the intervention. However, instrumental social support decreased at 6 and 12 months after the intervention.
Our 1-year attrition rate was high (35/90, 39%), as previously reported in similar longitudinal studies [
Our research found that videoconferencing effectively improved elderly residents’ depressive status at 3, 6, and 12 months. These results are consistent with a previous report [
We found that videoconferencing effectively improved elderly residents’ loneliness at 3, 6, and 12 months. These results are consistent with those of another study done in the United States [
Our research found that videoconferencing, a computer-mediated communication, had no effects on instrumental social support at the 3-month data collection time, as previously reported [
Our videoconference intervention also improved emotional social support at 3 and 12 months and appraisal support after 3 months. The effect of videoconferencing on appraisal and emotional support at 3 months is similar to our previous study [
From this point of view, we suggest that nursing home administrators increase the quality of communication by developing an interaction program such as arranging for family members to have a meal with residents at the nursing home and have a meal together via videoconference. One explanation for the long-term (12 months) decrease in emotional social support might be that nursing home residents feel safe or comforted by using videoconferencing as an alternative “social presence” so that they can immediately see their family member, even at a distance. Videoconferencing may offer them a chance to be part of family life. They also might feel comforted by seeing their family member’s actual state and would not worry that the family member was hiding a problem to allay anxiety if he or she could not visit [
The use of videoconference visits decreased over time. This decreased use of videoconferencing was likely due to a loss in the novelty of videoconferencing, lack of staff to help the residents operate the devices, and a need to remind family members to use videoconferencing (busy family members tended to forget to use videoconferencing). However, these possible reasons for decreased videoconference use need to be supported by further research. Furthermore, we found that videoconference use was high in some nursing homes, especially for those residents with relatives living overseas. However, our data were not significant due to the small sample size from each nursing home. We found no studies on the relationships between videoconference use and the characteristics of nursing home residents’ families. We therefore suggest that future research explore the relationships between videoconference use and characteristics of nursing home residents’ family members, factors influencing videoconference use, effects of videoconferencing on the health of elderly residents and their families, and the cost effectiveness of the videoconference program.
Although the experimental and control groups did not differ significantly in any dependent variable at any time point, the experimental group showed significant changes in depression, loneliness, and two social support measures over time compared with the control group after controlling for residents’ age and length of residency. The independent
Our research also showed that, after adjustment for residents’ age and length of residency, the time effects between the experimental and comparison groups remained the same for all outcome variables except emotional social support at 12 months. In other words, after adjustment for time and group effects, age and length of residency had almost no significant impact on all outcome variables, except emotional social support at 12 months.
Furthermore, the outcome variables of loneliness, lack of social support, and depression might have been associated with each other. However, when we analyzed each variable by GEE method with and without controlling for other variables, we found that the trends were not affected (data not shown). Further research is recommended to explore the associations among these variables and their possible impact on the time effects.
Our videoconferencing program was funded by the National Science Council, an agency of the Taiwan government (NSC97-2314-B-182-018), to Hsiu-Hsin Tsai, Principle Investigator. We thank Professor Hsiu-Hung Wang, Professor Hao-Hua Chu, and Associate Professor Hsu-Min Tseng for their contributions to the program development. We also thank Professor Yue-Cune Chang for statistical consultation.
None declared